What Musicians Actually Earn Per Stream

Pro-rata pooling has consequences people find counter-intuitive until they see them.

Your effective rate falls when total platform listening rises faster than platform revenue — more streams dividing the same pool. It rises when a service adds subscribers on a higher-priced tier. It differs enormously by territory, because a stream in a market with a $3 subscription contributes a fraction of what a stream in a $12 market contributes, and pooled accounting blends them.

That last point explains most of the confusion in artist forums. Two artists with identical stream counts can be paid very differently because one has an audience concentrated in high-ARPU markets and one does not. Neither was cheated; the accounting simply is not a price list.

There is a second model in play. Some services have moved to what is usually called artist-centric or user-centric allocation, where a subscriber's fee is divided among the artists that subscriber actually played rather than thrown into a global pot. It changes who wins without changing the total, and it tends to favour artists with committed listeners over artists with large passive playlist placement.

These are reported ranges for what reaches the rights holders — the owner of the recording, before any split with the artist. Treat them as the shape of the market rather than a rate card, and note the source column: some of these come from the platform's own published statements, the rest from distributor reporting.

One structural note that belongs with this table: the publishing side is set separately and statutorily. In the United States the Copyright Royalty Board fixes the mechanical rate for interactive streaming in five-year proceedings — the Phonorecords IV determination covers 2023 through 2027, with the headline pool percentage stepping up across the term. That determination is a Federal Register document with an explicit rate period, which makes it one of the few genuinely citable numbers in this entire field.

Money leaves the service in two streams that never touch each other, and confusing them is the single most common error in per-stream reporting.

The recording stream goes to whoever owns the master — a label, a distributor, or the artist. Out of every dollar the service pays out, somewhere around three-quarters of it flows here. The label then pays the artist a contractual royalty out of that, but only after recoupment, and recoupment is the whole ballgame.

The publishing stream goes to the composition: mechanical royalties administered in the US through The MLC at the CRB-set rate, and performance royalties through the PROs. That pool is far smaller — roughly a fifth of what the recording side receives — and it splits 50/50 between the publisher's share and the writer's share before any co-writer splits.

Which means a performing artist who wrote nothing sees only the recording line, and a writer who performed nothing sees only the publishing line, and the two of them will report wildly different per-stream figures for the identical song. Both are telling the truth.

Calculated at a blended recording payout of $0.0042 per stream, which sits in the middle of the reported bands for a typical mixed-territory audience. Substitute your own blended rate from a statement and the ratios hold.

An advance is a loan against future royalties, and almost every cost a label incurs on an artist's behalf gets charged back to the same account — recording budget, video production, tour support, sometimes a share of marketing.

The account recoups at the artist's royalty rate, not at the label's income. An artist on 18% who took a $250,000 advance needs the label to collect roughly $1.39 million in recording revenue before the artist sees a further cent, because only 18 cents of each collected dollar reduces the balance. The label is in profit long before the artist is paid anything.

This is not a scandal, it is the deal — the label carried the risk on a record that might have returned nothing. But it makes 'how much do you earn per stream' unanswerable for a large share of signed artists, because the honest answer is zero, and it will stay zero until an arithmetic threshold that the artist frequently has never been shown clears.

Worth knowing: the terms are not uniform any more. Several catalogues introduced legacy-artist unrecouped-balance forgiveness programmes for older deals, and some deals carry a reduced rate for the recoupment period and a higher rate after. Both change the table above materially, and both live in the contract rather than in any published rate.

A few mechanics that quietly reshape small-artist income and rarely appear in per-stream articles.

• ▸Minimum-stream thresholds. Spotify introduced a requirement that a track reach 1,000 streams in the prior twelve months before it generates recorded royalties at all, effective from early 2024. Tracks below the line generate nothing and the money stays in the pool for everyone else.
• ▸Fraud and artificial-streaming penalties. Distributors now pass through per-track charges when a service flags manipulated streams, which can turn a small payout negative.
• ▸Minimum payout balances. Most distributors hold funds until a threshold is met, so a statement showing earnings is not the same as money you can withdraw.
• ▸Territory withholding. Foreign withholding tax is deducted at source in several markets and reclaiming it requires the right treaty paperwork on file before the payment, not after.
• ▸Bundled-tier reallocation. When a service bundles music into a wider subscription, the share of that subscription allocated to music can fall, and every rate in the table above moves with it.

Apply the same test you would to any other money claim: what is the document, and when was it true?

A rate quoted from a platform's own published artist-economics report is dated and attributable. A statutory mechanical rate quoted from a Copyright Royalty Board determination carries an explicit rate period on its face. A figure quoted from an artist's screenshot of one month's dashboard is real but describes one audience in one month. A figure quoted from another article quoting another article describes nothing.

And any per-stream number presented without the deal structure attached is incomplete by construction. The platform's payout and the artist's payment are separated by a contract, and the contract is where almost all the variance lives.